Amniotic Fluid Interleukin 6 in Preterm Labor
Association with Infection
Roberto Romero, Cecilia Avila, Uma Santhanam,* and Pravinkumar B. Sehgal*
Department ofObstetrics and Gynecology, Yale University School ofMedicine, New Haven, Connecticut 06510;
and *The Rockefeller University, New York 10021
To evaluate whether IL-6 participates in the host response to
intrauterine infection, we studied IL-6 bioactivity and isoforms
in amniotic fluid (AF). Two different assays for IL-6 were used:
the hepatocyte stimulating factor assay (in Hep3B2 cells) and
the SDS-PAGE/immunoblot assay. IL-6 determinations were
performed in 205 AF samples. Samples were obtained from
patients in the midtrimester of pregnancy (n = 25), at term
with no labor (n = 31), at term in active labor (n = 40), and
from patients in preterm labor (n = 109). Higher AF IL-6
levels were observed in women in preterm labor with intraam-
niotic infection than in women in preterm labor without in-
traamniotic infection (median = 375 ng/ml, range = 30-5000
ng/ml vs. median = 1.5 ng/ml, range = 0-500, respectively, P
< 0.0001). The 23-25- and 28-30-kD IL-6 species could be
readily detected in SDS-PAGE immunoblots performed di-
rectly on 10-,ul aliquots ofAF from patients with intraamniotic
infection. Among women in preterm labor with culture-nega-
tive AF, those who failed to respond to subsequent tocolytic
treatment had higher AF IL-6 concentrations than those who
responded to therapy (median = 50 ng/ml vs. median = 1.2
ng/ml, respectively, P < 0.05). Only low levels of IL-6 were
detected in AF obtained from normal women in the midtrimes-
ter and third trimester of pregnancy. Decidual tissue explants
obtained from the placentas of women undergoing elective ce-
sarean section at term without labor (n = 11) produced IL-6 in
response to bacterial endotoxin. In a pilot study, AF IL-6 was
determined in 56 consecutive women admitted with preterm
labor. All patients (n = 10) with elevated AF IL-6 (cutoff = 46
ng/ml) delivered a premature neonate. 4 of these 10 patients
had positive AF cultures for microorganisms. These studies
implicate IL-6 in the host response to intrauterine infection
and suggest that evaluation of AF IL-6 levels may have diag-
nostic and prognostic value in the management of women in
preterm labor. (J. Clin. Invest. 1990. 85:1392-1400.) prema-
turity-parturition- decidua * cytokines-bacterial endotoxins
Prematurity is the leading cause of perinatal mortality and
morbidity worldwide (1). A growing body ofevidence suggests
an association between subclinical intrauterine infection and
Address correspondence to Dr. P. B. Sehgal, The Rockefeller Univer-
sity, 1230 York Avenue, New York, NY 10021.
Receivedfor publication 29 June 1989 and in revisedform 7 De-
preterm labor (PTL)' (reviewed in reference 2). Recently, we
have estimated that at least one of every five premature neo-
nates is born to a mother with an intraamniotic infection (3,
4). The majority ofthese infections are subclinical and difficult
to diagnose (3, 4). The participation ofcytokines in the inflam-
matory process associated with intraamniotic infection is sup-
ported by the detection of IL-1 and tumor necrosis factor
(TNF) in the amniotic fluid (AF) ofwomen with this condition
(5, 6). The purpose ofthis study was to evaluate the participa-
tion of IL-6 in the pathophysiology of premature labor.
IL-6 has been implicated as a major mediator of the host
response to infection and tissue damage (7, 8). This cytokine
consists of a group of differentially modified phosphoglyco-
proteins ranging in size from 23 to 30 kD and additional com-
plexes of 43-45 kD (8-13) derived from a single gene located
at 7p2l in the human genome (14-16). IL-6 gene expression is
induced by several inflammation-associated cytokines (in-
cluding IL- 1, TNF, and interferons), bacterial products, RNA-
and DNA-containing viruses, and second messenger agonists
(diacylglycerol, cAMP, and Ca2+) that activate any ofthe three
major signal transduction pathways (17-26). Cell types capa-
ble of secreting IL-6 in response to stimulation include fibro-
blasts, monocytes/macrophages, endothelial cells, keratino-
cytes, and endometrial stromal cells (7-27).
IL-6 has a broad range of biological effects (reviewed in
reference 25). It elicits major changes in the biochemical, phys-
iological, and immunological status ofthe host (e.g., the "acute
phase" plasma protein response, activation ofT and NK cells,
and stimulation ofproliferation and immunoglobulin produc-
tion by B cells). The alterations in plasma protein composition
mediated by IL-6 are thought to seal the site of tissue injury
and reduce the systemic effects ofinfection and tissue damage
(7, 25). For example, increased concentrations of fibrinogen
and other coagulation factors may promote thrombus forma-
The acute phase plasma protein response may be impor-
tant in the context of intraamniotic infection. Clinical studies
have indicated that elevation of maternal serum C-reactive
protein (CRP) often precedes the development of clinical
chorioamnionitis and the onset of premature labor in women
with preterm premature rupture of membranes (28-30). Fur-
thermore, patients in PTL with elevated levels of CRP are
more likely to be unresponsive to tocolytic therapy than those
with nondetectable CRP (30-32). As IL-6 plays a critical role
in the induction of CRP synthesis (33, 34), it was considered
likely that this cytokine participated in the host response to
1. Abbreviations used in this paper: ACT, a1-antichymotrypsin; AF,
amniotic fluid; CRP, C-reactive protein; HSF, hepatocyte stimulating
factor; PTL, preterm labor; r, recombinant; TL, term labor; TNF,
tumor necrosis factor.
R. Romero, C. Avila, U. Santhanam, and P. B.Sehgal
J. Clin. Invest.
©The American Society for Clinical Investigation, Inc.
Volume 85, May 1990, 1392-1400
IL-6 bioactivity was measured and IL-6 isoforms were
characterized in the AF ofwomen with and without intraam-
niotic infection. Highly elevated AF IL-6 levels were observed
in women with PTL and intraamniotic infection who pro-
gressed to delivery. AF samples from this group of patients
contained the highest levels of IL-6 observed by us in any
human body fluid. Indeed, the levels were so high (median
= 375 ng/ml; range = 30-5,000 ng/ml) that the 23-25- and
28-30-kD IL-6 species could be readily detected in immuno-
blots performed directly on 10-jd aliquots ofthe fluid. Women
with PTL unresponsive to tocolysis had a higher median AF
concentration of IL-6 than women who responded to tocol-
ysis. These observations were confirmed in a pilot study of 56
consecutive women admitted with PTL. All patients with ele-
vated AF IL-6 (cutoff = 46 ng/ml) subsequently delivered a
premature neonate. Furthermore, all patients with culture-
positive AF had elevated IL-6 in their AF as defined in this
assay. These data suggest that elevated AF IL-6 levels may be
of value in the detection of intraamniotic infection and may
also provide prognostic information regarding the effective-
ness oftocolysis in impeding preterm delivery.
Patient population and collection ofAF. AF was collected from women
in four different populations: group 1, women admitted with PTL and
intact membranes on whom amniocenteses were performed for micro-
biological assessment ofthe amniotic cavity and for fetal lung maturity
studies (n = 109); group 2, women in the midtrimester of pregnancy
(gestational age: 16-18 wk) who underwent amniocenteses for genetic
indications (maternal age > 35 yr) (n = 25); group 3, women in the
third trimester (gestational age: 37-40 wk) who had amniocenteses for
the assessment offetal lung maturity before elective cesarean section (n
= 31); and group 4, women in spontaneous active labor (cervical
dilatation of at least 6 cm) at 38-41 wk (term labor [TL]) from whom
AF was obtained by transvaginal amniotomy (n = 40).
To study the relationship between intraamniotic infection, PTL,
and AF IL-6 concentrations, a cross-sectional study of patients in
group 1 (PTL) was constructed. Patients with PTL were divided into
three subgroups according to their response to tocolysis and the results
of their AF culture: subgroup la, women with PTL and negative AF
cultures who were responsive to tocolysis; subgroup lb, women with
PTL and negative AF cultures who were unresponsive to tocolysis and
delivered a preterm neonate; and subgroup 1c, women with PTL and
intraamniotic infection who delivered a preterm neonate. In addition,
we also conducted a pilot study to explore the potential clinical value
ofAF IL-6 determinations. This cytokine was assayed in the AF of 56
consecutive women admitted with the diagnosis of PTL.
An intraamniotic infection was considered to be present ifmicrobi-
ological studies ofAF yielded a positive result (2). PTL was defined as
the presence ofregular uterine contractions with a frequency ofat least
two every 10 min for at least 60 min. A 132-adrenergic drug (ritodrine in
most cases) was administered intravenously as the tocolytic agent ac-
cording to a protocol described elsewhere (35). Amniocentesis was
performed before the initiation of therapy. Failure of tocolysis was
diagnosed when cervical dilatation progressed beyond 5 cm and/or
After defining the criteria for entry into the study, AF samples were
obtained from the AFbank in the Department ofObstetrics and Gyne-
cology, Yale University School of Medicine (New Haven, CT). This
bank consists of aliquots ofAF that were not used for clinical testing.
AF samples were centrifuged at 400 g for 10 min at 4°C immediately
after collection, and the fluid was separated from the cell pellet and
stored frozen at -70°C. AF samples analyzed for IL-6 in this study
were from consecutive patients admitted to the hospital who fulfilled
the criteria ofeach ofthe study groups. Informed consent was obtained
from all patients following the guidelines of the Yale University
Human Investigations Committee.
Microbiological studies. AF in a capped plastic syringe was trans-
ported to the microbiology laboratory immediately aftercollection and
plated within 30 min of amniocentesis. AF was cultured for aerobic
and anaerobic bacteria as well as for Mycoplasma hominis and Urea-
plasma urealyticum, as described previously (36). AF samples from
women in groups 2 and 3 were sterile. AF cultures were not performed
on samples obtained by transvaginal amniotomy (group 4). Previous
experience has indicated that the results ofthese cultures do not reflect
the microbiological status of the amniotic cavity because contamina-
tion with cervical or vaginal flora cannot always be avoided during
transvaginal amniotomy (2).
Cultures of decidual tissue explants. Decidual explants were ob-
tained from 11 pregnant women at term who had experienced un-
complicated pregnancies and were scheduled to have elective repeat
cesarean sections. None had spontaneous onset oflabor before surgery.
Care was taken to avoid contact ofthe placenta with the skin or other
potential sources of endotoxin. Immediately after delivery of the pla-
centa, the decidua adherent to the chorion was mechanically stripped
with sterile forceps. The explants were placed in ice-cold, pyrogen-free
Krebs-Ringer's solution. They were incubated inDME (Gibco Labora-
tories, Grand Island, NY) with 10% FCS, 100 U/ml of penicillin,
streptomycin (final concentration 100gg/ml),and 4AML-glutamine.
Explants were -1 cm X
media in 35-mm plastic tissue culture dishes (Coming Glass Works,
Corning, NY). Explants were incubated at 37"C in a humidified incu-
bator with 95% 02 and 5% CO2 for 20 h. The medium was centrifuged
at 200 g for 10 min. Supernatants were removed and frozen immedi-
ately at -20'C until the time ofthe assay.
Hepatocyte stimulatingfactor(HSF) assayforIL-6 biologicalactiv-
ity. The levels ofbiologically active IL-6 in AF samples were estimated
using the HSF assay as previously described in detail (9, 10, 12, 13).
Briefly, the human hepatoma cell line Hep3B clone 2, obtained from
the American Type Culture Collection (Cat. No. HB8064; Rockville,
MD), was grown to near-confluence in 24-well tissue culture plates
(Falcon Plastics, Cockeysville, MD). Appropriately diluted AFsamples
(diluted 1:5-1:250) were added to Hep3B2 cultures (0.5 ml/well) in the
presence ofexcess insulin (8 ,ug/ml) and dexamethasone (1AM).After
incubation for 20-24 h at 37°C, the cultures were washed with PBS
and incubated for an additional 24 h in methionine-free medium con-
taining [35S]methionine (100ACi/ml;DuPont-New England Nuclear,
Boston, MA). The culture medium from each well was then collected,
and the amount of labeled a1-antichymotrypsin (ACT) secreted was
quantitated by immunoprecipitation (rabbit antiserum to human ACT
was purchased from Calbiochem-Behring Corp., San Diego, CA),
SDS-PAGE, autoradiography, and densitometry (Ultroscan XL laser
densitometer; LKB Instruments, Inc., Gaithersburg, MD) using pro-
cedures described earlier (10, 12, 13). A purified preparation ofnatural
human IL-6 (produced by IL-la-induced fibroblasts) was used as an
internal laboratory standard, and appropriate dilutions ofthis prepara-
tion were also assayed in duplicate in every experiment. The concen-
tration ofIL-6 in the laboratory standard (2Ag/ml)was determined by
silver-staining of the IL-6 proteins after SDS-PAGE, and the same
value (2Ag/ml)was also arrived at independently by comparing this
mixture with electrophoretically homogeneous Coomassie blue-
stained Escherichia coli-derived recombinant (r) IL-6 (8) in immuno-
blots using rabbit anti-rIL-6 (8). Concentrations ofbiologically active
IL-6 in AF are expressed in nanograms/milliliter by comparison to this
The biological activity in AF samples observed in this HSF assay
was verified to be due to IL-6, based on the ability ofa rabbit polyclo-
nal antiserum (1:100 dilution) prepared against purified E. coli-de-
rived human IL-6 (8) to completely neutralize the activity. The proper-
ties and characteristics of this anti-rIL-6 antiserum have been exten-
sively described previously (8-13, 37, 38). This antiserum specifically
blocks IL-6 activity not only in hepatocyte stimulation assays but also
1 cm in size and were incubated in 2 ml of
Amniotic Fluid Interleukin 6
in B cell growth and differentiation assays, hybridoma growth assays,
and T cell activation assays. We have verified that IL- 1a at concentra-
tions up to 50 ng/ml, IL-1#at 2 ng/ml, TNF at concentrations up to 1
ug/ml, and bacterial LPS up to 5Ag/mldo not affect ACT synthesis in
the Hep3B2 cells used (8, 10, and data not shown). Nevertheless, AF
samples in group 1 (PTL) were also assayed in the presence of excess
neutralizing antibodies to both IL- 1a and IL-I,8 (2 ug/ml each), be-
cause some of these AF samples were previously found to contain
immunoreactive IL- 1 fI (up to a maximum of2-4 ng/ml; Romero, R.,
unpublished observations). The murine neutralizing monoclonal anti-
bodies to human IL-Ia or IL-1,8 were kindly provided by Dr. A. C.
Allison (Syntex Research, Palo Alto, CA). At 1gg/ml,both antibodies
block 1 ng/ml of IL- la or IL- I#in the thymocyte comitogenic assay
and the fibroblast proliferation assay (39). An anti-TNF monoclonal
antibody was obtained from the Suntory Institute for Biomedical Re-
search (Osaka, Japan);
rTNF activity in the L929 cytotoxicity assay.
In summary, the HSF assay for biologically active IL-6 consists of
two parts: (a) estimation of the enhancement of ACT synthesis in
Hep3B2 cells, and (b) inhibition ofthis enhancement by anti-rIL-6. A
major advantage ofthis assay is that it is not affected by the presence of
bacterial products in the test samples.
Immunoblot assayfor IL-6 proteins. AF samples from women with
PTL and intact membranes (group I) were also assayed for IL-6 con-
tent using an immunoblot procedure (8-10, 12, 13). 10-Ad aliquots of
AF were electrophoresed through SDS-PAGE (17.5%) under reducing
and denaturing conditions, electroblotted onto nitrocellulose paper,
and probed using rabbit anti-rIL-6 (8-13) and the ABC Elite Vecta-
stain kit (Vector Laboratories, Inc., Burlingame, CA). Each blot also
contained appropriate dilutions ofthe laboratory standard for natural
IL-6. As controls, additional blots were probed using rabbit preim-
mune serum, serum from an unrelated rabbit, or the first antibody
omitted from the reaction.
Statistical analysis. Comparisons of IL-6 estimates in AF samples
from different groups were conducted using a Kruskal-Wallis one-way
analysis of variance. The Dunn test was used for post-hoc multiple
comparisons among groups (True Epistat; Epistat Services, Richard-
I ug/ml ofthis antibody neutralizes 2 ng/ml of
We have evaluated IL-6 bioactivity in AF from women at
midtrimester, at term not in labor, at term in active spontane-
ous labor, and from patients in PTL with and without in-
traamniotic infection. We shall focus on the results ofAF IL-6
from women with PTL because these patients presented with
intact membranes, and, therefore, a positive AF culture is a
reliable indicator ofmicrobial invasion ofthe amniotic cavity.
IL-6 in AF in PTL. Fig.
IL-6/HSF assay carried out on AF samples from women in
PTL. A 1:5 dilution ofAF from six patients in each ofthe three
subgroups Ia, lb, and Ic (as defined in Methods), is illustrated.
As can be seen, the IL-6/HSF activity of AF samples from
women in PTL who were responsive to tocolysis and who were
AF culture-negative (subgroup
samples of women in PTL who were not responsive to tocol-
ysis and who were AF culture-positive (subgroup Ic) is the
highest (also see Figs. 3 and 4). It is striking that all ofthe AF
samples in subgroup lc (Figs.
positive for IL-6/HSF activity.
That the biological activity observed in this IL-6/HSF
assay in these and other AF samples was due specifically to
IL-6 was verified by a separate set of experiments similar to
those summarized in Fig. 2. Appropriate dilutions ofAF (three
each from Fig. 1, B and Cand three from a separate group with
1 summarizes a representative
1a) is the lowest. That of AF
1 C, 3, and 4) were strongly
in the AF ofwomen with PTL.
AF diluted 1:5(500Ad final
volume) from women in PTL
was assayed for its ability to
ACT inHep3B2cultures. The
figureis acompositeof data in
the stimulation ofACTsyn-
samplesfrom six differentpa-
tients in each of thefollowing
threesubgroups: A, AF cul-
sive to tocolysis (subgroup la);
PTL notresponsiveto tocol-
ysis (subgroupIb);and C, AF
culture-positiveand PTL not
responsive to tocolysis (sub-
group Ic). For comparison, D
illustrates ACT stimulationby
the natural IL-6 standardprep-
aration at 10, 5, 2, 1, and 0
sets 1, 2, 3, 4, and 5, respec-
1 2 3 4 5 6
PTL and intraamniotic infection) were mixed with a 1:100
dilution of the anti-rIL-6 antibody (lanes a),and the residual
HSF activity was assayed. The data shown inFig.2clearly
demonstrate that this anti-rIL-6antibodywas able tostrongly
inhibit the HSF activityobserved. Additionally,we have veri-
tivity byanti-rIL-6 anti-
andB)in which the
samples (500 IAdfinal
tures wasassayedafter incubation with rabbit anti-rIL-6 antiserum at
1:100 dilution(lanes a).Three AFsamplesfrom each ofthe follow-
ingthreegroupswere evaluated: A2, B5, B6,AFculture-negative,
and PTL notresponsivetotocolysis (subgroup
plesare fromamongthe six illustrated inFig.
culture-positive,and PTL notresponsivetotocolysis (subgroup 1c;
these threesamplesare fromamongthe six illustrated inFig. 1 C);
A4, B8, B9correspondto aseparateset ofsamplesfrom women with
PTL and infection. In additionalexperiments,control unrelated rab-
bit serum at 1:100 dilution had little effect on the stimulation of
ACTsynthesis (not shown).Residual HSFactivityseen in some of
the lanes(particularly 7a),even in thepresenceofanti-rIL-6 at 1:100
dilution,is consistent with thevery highconcentrations of IL-6 in
thesesamples (200-5,000 ng/mlassubsequentlyverified byimmun-
oblotassays;seeFig. 3). bl, ACTsynthesisin unstimulated control
1b;these three sam-
IB); Al, A3, B7,AF
R. Romero, C. Avila, U. Santhanam, and P. B. Sehgal
fied in similar neutralization experiments that the following
reagents have little or no effect on the HSF activity observed in
AF samples: (a) preimmune or unrelated rabbit serum, (b)
neutralizing anti-TNF antibody, and (c) a combination ofthe
neutralizing anti-IL-la and anti-IL-l1B antibodies. We thus
conclude that the observed HSF biological activity in AF sam-
ples can be ascribed to IL-6.
To substantiate this conclusion further, we have evaluated
the IL-6 content of AF samples from women in PTL by im-
munoblot analysis. This procedure allows characterization of
the IL-6 species present in AF. Fig. 3 illustrates immunoblot
data obtained from 10 samples each from subgroups la and
lc. It is clear that AF samples in subgroup lc (PTL, not re-
sponsive to tocolysis and AF culture-positive) contain IL-6 at
such high concentrations that the 23-25- and 28-30-kD spe-
cies of this cytokine can be readily detected using only a10-p4l
aliquot in this assay. The data in Fig. 3 are consistent with
those in Fig. 1 in that the lowest IL-6 concentrations are seen
in AF samples in subgroup la, and the highest levels in sub-
group lc. Immunoblot assays on the AF samples in group 1
were consistent with data obtained using the HSF bioassay
(data not shown). Strikingly, all ofthe samples in subgroup 1 c
are strongly positive for IL-6 in both the HSF and immunoblot
assays (also see below).
The immunoblot data in Fig. 3 provide the first description
of IL-6 species present in human AF. These data show that
differentially modified IL-6 species (the 23-25-kD species are
O-glycosylated; the 28-30-kD species are 0- and N-glycosy-
lated [8, 10, 25]) are indeed present in human body fluids.
These immunoblot assays reveal that IL-6 isoforms are present
in AF at concentrations as high as 1-5 gg/ml.
The concentrations of IL-6 in all the AF samples from
women in PTL (group 1) were estimated using the HSF assay.
AF samples were diluted as appropriate (up to 1:250) to obtain
a stimulation in ACT synthesis that would be within the log-
linear range of the assay. Furthermore, strongly positive sam-
ples (in groups lb and
excess neutralizing antibody to both IL- 1 a and IL- 1f. Fig. 4 a
summarizes the estimates of IL-6 content in all of the AF
samples in the group with PTL. Patients with intraamniotic
infection had significantly greater concentrations of AF IL-6
than women without intraamniotic infection regardless of
their response to tocolysis (median = 375 ng/ml, range
= 30-5,000 ng/ml vs. median = 1.5 ng/ml, range = 0-500,
respectively, P < 0.0001). Patients without intraamniotic in-
fection but refractory to tocolytic therapy had higher median
concentrations of AF IL-6 than those women without in-
1c) were reassayed in the presence of
traamniotic infection who responded to tocolysis (P < 0.05)
(Fig. 4 b).
The clinical picture of several outliers is noteworthy. One
patient with PTL responsive to tocolysis had an AF IL-6 level
of 125 ng/ml. This patient presented with PTL and a tempera-
ture of 100°F. Her white blood cell count was 12,200 with a
differential count of 81% segmented neutrophils and 5%
monocytes. Although AF and urine cultures were negative, it
is possible that this patient had a self-limiting inflammatory
reaction that could not be diagnosed. She delivered at term 16
Two women in the group of patients with PTL who were
unresponsive to tocolysis (group lb) had AF IL-6 concentra-
tions of 500 ng/ml. AF from one of these patients stained
positive for gram-positive cocci and gram-negative rods, but
the culture grew only Corynebacterium species. Isolation of
this microorganism is generally interpreted as consistent with a
skin contaminant. This interpretation may have been errone-
ous. The presence of elevated IL-6 suggests that there was an
ongoing inflammatory reaction in the amniotic cavity. The
second patient was clinically suspected to have chorioamnion-
itis because of the association of fever and PTL. Although the
gram stain ofAF was negative for bacteria, and the culture was
negative, histopathologic examination ofthe placenta showed
intense chorioamnionitis. Therefore, this patient may have
had an intraamniotic infection that escaped detection using
current microbiologic techniques, an infection limited to the
extraamniotic membranes, or an inflammatory reaction of a
noninfectious etiology. In any case, the high AF IL-6 levels
appear to be indicative of an ongoing inflammatory process.
Elevated AF IL-6 levels are associated with intraamniotic
infection by a variety ofmicroorganisms including gram-posi-
tive and gram-negative species and Ureaplasma urealyticum
(see Table I). These data strongly suggest that elevated AF IL-6
levels are indicative of an inflammatory reaction associated
with infection regardless of the specific microorganism in-
IL-6 in the AFfrom spontaneous TL. We extended our
studies of AF IL-6 levels during parturition to a group of pa-
tients in spontaneous TL (group 4). As controls, we studied AF
IL-6 levels in patients in midtrimester (group 2) and in women
at term but not in labor (group 3). Fig. 5 summarizes the data
obtained. AF from women in the midtrimester and third tri-
mester (at term) of pregnancy contained detectable but low
levels of IL-6 (median 10 ng/ml in group 2 and 13 ng/ml in
group 3). However, AF IL-6 levels were higher in women in
spontaneous labor at term (group 4) than in women who were
Figure3. Multiple forms of IL-6proteinsin AF in
PTL. Thiscomposite figure illustrates SDS-PAGE/
immunoblot assays for IL-6 carried out using l0-tl
AF aliquots (undiluted final volume) from 10pa-
tients ineach of the following two groups: A, AF cul-
ture-negative and PTL responsive to tocolysis (sub-
group la, two ofthe samples are from among the six
PTL not responsive to tocolysis (subgroup Ic, four of
the samplesare fromamongthe six illustrated inFig.
1C). Lanes a, b, and c in each panel illustrate im-
illustrated in Fig.
1A); B, AF culture-positive and
munoblot assays using the natural IL-6 standard preparation at 4, 2, and 1 ng of antigen per lane in A and at 10, 2, and 0.4 ng per lane in B. In
additional immunoblot analyses ofAF samples, unrelated rabbit serum or the second anti-rabbit antibody by itself did not react with the
23-25- and 28-30-kD anti-rlL-6-immunoreactive proteins (data notshown).
Amniotic Fluid Interleukin 61395
Figure 4. AF IL-6 levels
in women with PTL. a,
Scatter diagram illus-
trating AF IL-6 levels in
three different sub-
groups of patients: A,
AF culture-negative and
responsive to tocolysis
(subgroup la,n = 25;
median IL-6 = 1.2
negative and not re-
sponsive to tocolysis
(subgroup Ib, n = 13;
ng/ml, range = 0-500
ng/ml); C, AF culture-
positive and not respon-
sive to tocolysis (sub-
~group Ic, n=15; me-
n = 14
.dian IL-6 = 375 ng/ml,
range = 30-5,000
H = 30.425; P
=0.0000002; A compared with B, P < 0.05; B compared with C, P
< 0.05; C compared with A, P < 0.05 (Dunn's test). b, Replot of
data for subgroups A and B shown in a.
not in labor at term (group 3) (median = 19 ng/ml, range
= 4-500 for women in active labor vs. median = 13 ng/ml,
range = 3.5-60 ng/ml; P < 0.01). The major difference be-
tween these two groups can be attributed to several patients in
active labor who had marked elevations ofAF IL-6 levels. A
subclinical intraamniotic infection is the most likely explana-
tion. Indeed, previous histopathologic studies indicate that up
to 10% of normal term placentas have inflammatory lesions
consistent with the presence ofan intraamniotic infection (32).
Bacterial LPS induces IL-6 production by decidual tissue
explants. We have tested whether decidual tissue explants
from normal placentas can be induced to secrete IL-6 by bacte-
rial LPS. Fig. 6 summarizes an evaluation of IL-6 bioactivity
in the medium ofuninduced and LPS-induced decidual tissue
explant cultures prepared from the placentas of 11 different
women. Tissue from each individual placenta was cultured in
the absence (lanes u) or the presence (lanes 1) ofbacterial LPS
(25 ng/ml) for 24 h. Fig. 6 illustrates that several of the deci-
dual tissue explants spontaneously secrete significant amounts
ofIL-6 and that bacterial endotoxin increased IL-6 production
by all explants. In additional assays, we have verified that
anti-rIL-6 strongly inhibits the IL-6 activity in LPS-induced
samples illustrated in Fig. 6. These data are consistent with the
hypothesis that bacterial products can stimulate tissues in the
maternal-fetal interface to produce IL-6.
Clinical value ofAF IL-6 measurements. We designed a
pilot study to explore the potential value ofroutine IL-6 mea-
surements in clinical practice. AF samples from 56 consecu-
tive women with the clinical diagnosis ofPTL were tested for
IL-6 content. Of these patients, 25 subsequently delivered a
preterm neonate and 31 delivered at term. IL-6 content was
assayed in this pilot study using the HSF assay, in which the
stimulation of[35S]methionine-labeled ACT was monitored by
immunoprecipitation, PAGE, and autoradiography. Visual
inspection ofthis autoradiogram allowed us to readily identify
10 AF samples that had clearly elevated IL-6 activity. Subse-
quent quantitation revealed that this subgroup of 10 fluids had
AF IL-6 levels that were equal to or greater than 46 ng/ml. We
then tabulated the clinical outcome in these 10 patients; all
these patients delivered preterm neonates. Clinical informa-
tion, microbiological results, and placental pathology of these
patients are displayed in Table II. It is noteworthy that the four
patients with positive AF cultures for microorganisms also had
elevated AF IL-6. Furthermore, several of the patients who
had elevated AF IL-6 but negative AF cultures had histopatho-
logic evidence ofchorioamnionitis.
Table I. Microbiological Data and IL-6 in Women with PTL and Intact Membranes
Group B Streptococcus
Group B Streptococcus
Mixed anaerobic flora
1396 R. Romero, C. Avila, U. Santhanam, and P. B. Sehgal
Figure5. AF IL-6 levels
in women in midtri-
mester and at term.
Scatter diagram illus-
trating AF IL-6 levels in
three different groups of
patients: A, midtrimes-
ter (group 2, n= 25;
smedian IL-6 = 10
ng/ml); B,at term but
not in labor(group 3, n
= 31; median IL-6 = 13, range = 0-60 ng/ml); and C, at term and in
spontaneous labor (group 4, n = 40; median IL-6 = 19.5 ng/ml,
range = 4-500 ng/ml). Kruskal-Wallis H = 23.35; P = 0.0000085; A
compared with B, P > 0.05; B compared with C, P < 0.05; C com-
pared with A, P < 0.05 (Dunn's test).
This is the first report to describe the presence of IL-6 in AF.
We describe a dramatic increase in the AF IL-6 levels in
women with intrauterine infection and PTL. The data ob-
tained suggest that AF IL-6 levels may be of diagnostic and
prognostic value in the management ofPTL.
Biologically active IL-6 was detected in AF using the HSF
assay in Hep3B2 cells; this bioactivity was confirmed to be due
to IL-6 per se by its neutralization with an anti-rIL-6 anti-
serum. Although the HSF assay is not as sensitive as other
bioassays available for IL-6 (e.g., the hybridoma growth factor
assay) (25), it is robust and insensitive to the presence ofbacte-
rial products in the body fluid tested. Immunoblot analyses
confirmed the presence ofthe 23-25- and 28-30-kD IL-6 spe-
cies in AF. This extends previous descriptions ofthe heteroge-
neity of IL-6 species in human serum/plasma, cerebrospinal
fluid, and synovial fluid (10, 12, 13, 25) to the AF. IL-6 hetero-
geneity is, thus, a very general feature of this cytokine in
human body fluids and is due to N- and O-glycosylation and
differential phosphorylation of the 23-30-kD IL-6 species (8,
9, 25). The 23-25-kD IL-6 species are O-glycosylated, whereas
the 28-30-kD IL-6 species are both 0- and N-glycosylated.
Additionally, higher molecular mass immunoreactive com-
plexes (43-45-kD) have been reported in serum/plasma and in
synovial fluid (10-13). In cell culture, induced human fibro-
blasts, monocytes, endothelial cells, keratinocytes, and endo-
metrial stromal cells have all been shown to secrete multiple
differentially modified IL-6 species of molecular mass 23-30
kD in cell culture (8, 9, 11, 25). The biological consequences of
this heterogeneity are unclear.
A major finding in the present study is that the AF of
women with PTL and intraamniotic infection contained very
high levels of IL-6. All of the AF samples from'women with
PTL and infection were strongly positive for IL-6. This indi-
cates that IL-6 is a participant in the host response to intraam-
niotic infection. Microbiologic data from these patients dem-
onstrate that elevated AF IL-6 levels are observed in women
with intraamniotic infections due to a wide variety of organ-
isms including gram-negative and gram-positive bacterial spe-
It should be stressed that PTL leading to preterm delivery
in the absence ofdetectable infection was also associated with
elevations in AF IL-6 levels, although of a lesser magnitude
than that observed in cases ofintraamniotic infection. Several
possible explanations for this observation may be considered.
First, this subgroup of women may have had an intrauterine
inflammatory reaction unrelated to intraamniotic infection
(i.e., an extraamniotic infection or a noninfectious inflamma-
tory process). Second, an elevation of IL-6 may be associated
with the physiologic process ofparturition. Third, an intraam-
niotic infection may have escaped detection using standard
microbiological techniques. This latter possibility is a likely
explanation for two specific cases with elevated AF IL-6 con-
centrations (500 ng/ml). The first case had a positive gram
stain of AF for bacteria, but the AF grew Corynebacterium
species, which are considered a skin contaminant. The second
case had clinical and histopathologic evidence of chorioam-
nionitis, but no bacteria could be recovered. Further studies
are required to clarify these issues.
A perennial clinical problem is the interpretation ofmicro-
biological cultures ofbiological fluids in which the initial gram
stain and culture yield conflicting results. A potential role for
IL-6 in clinical practice could be to help identify between sam-
ple contamination or culture failure. For example, a positive
gram stain ofAF with elevated IL-6 levels is more likely to be
due to a true intraamniotic infection even with a negative AF
culture. Additionally, in one of the cases studied by us, it is
unclear whether the Corynebacterium isolated was a skin con-
taminant introduced into the sample during the amniocentesis
procedure or a true pathogen. The identification of microor-
ganisms in gram stains coupled with a high level ofIL-6 in AF
suggests a pathogenic role for microorganisms. This concept
broadens the potential clinical utility of cytokine analysis of
In the context ofPTL, it is ofconsiderable importance that
women who went on to deliver a premature neonate had ele-
vated levels ofAF IL-6, regardless ofthe presence orabsence of
documented infection. This observation may have important
clinical applications in obstetrics. It would seem that a high AF
IL-6 level may identify a group of patients who would not
benefit from tocolysis. Tocolytic therapy is associated with
Figure 6. Production of IL-6 by decidual tissue explants in response
to LPS. Decidual tissue explants from 11 placentas from women un-
dergoing elective cesarean sections without spontaneous labor were
cultured overnight in the absence (u) or presence ofLPS (25 ng/ml)
(1), and the accumulation of IL-6 bioactivity in the culture medium
was monitored (A). B illustrates the stimulation ofACT synthesis in
duplicate cultures by the natural IL-6 standard preparation used at
10, 2.5, 1, 0.4, and 0 ng/ml in this experiment (sets 1-5).
Amniotic Fluid Interleukin 6
Table II. Clinical Data ofWomen with ElevatedAFIL-6 (from a Pilot Study of56 Consecutive Patients Admitted with PTL)
Patient had been previously treated
Patient had been previously treated.
The neonate died ofgroup B
significant side effects for both mother and fetus (32, 40). Our
findings (also see below) justify additional larger prospective
studies to determine whether evaluation ofAF IL-6 levels can
help identify patients in PTL who will progress to delivery
despite tocolytic treatment.
The association between AF IL-6 levels and parturition
was also evaluated at term. We observed that the median con-
centration of AF IL-6 was higher in women in spontaneous
labor at term than in women at term who were not in labor.
These observations are similar to those reported by us earlier
for AF IL-1 bioactivity (5). Inspection of the IL-1 and IL-6
data sets reveals that this difference is due to a subset of pa-
tients in active labor who had elevations ofboth IL- I and IL-6
in their AF samples. It is possible that this subset consists of
patients with subclinical intraamniotic infection or chorioam-
nionitis. Because the present study did not include histopatho-
logic examination of placentas from women delivering at
term, we cannot directly address this question. Nevertheless,
our observations provide a basis for constructing a prospective
study to explore this question. A limitation ofour study is that
the method used to collect AF in women in active labor at
term was different from that used to retrieve fluid from women
at term not in labor (transvaginal amniotomy vs. transabdom-
After obtaining data that strongly linked elevated AF IL-6
levels to, first, intraamniotic infection and, second, parturi-
tion, we designed a pilot study to evaluate the clinical value of
AF IL-6 measurements in women with PTL. The most striking
observation was that all patients with elevated AF IL-6 went
on to deliver a premature neonate. Four ofthese patients had
positive AF cultures, implicating an intraamniotic infection as
the etiologic factor responsible for preterm delivery and also
for an elevation ofAF IL-6. On the other hand, a demonstra-
ble intraamniotic infection (defined as a positive AF culture
for microorganisms) was absent in the remaining six patients
with detectable AF IL-6. However, histological signs of cho-
rioamnionitis were demonstrated in three ofthe three cases in
which placentas were available for examination. This suggests
that an intrauterine infection was present but may have eluded
detection with microbiological techniques. This is probably
the case in two patients who had been treated with antibiotics
before amniocentesis. Antibiotic treatment may have ham-
pered our ability to recover microorganisms but not our ability
to detect IL-6 as an index ofthe host response to infection.
IL-6 in the AF is likely to be of both maternal and fetal
origin. We have recently shown that freshly explanted endo-
metrial stromal cells, a normal component of the maternal
decidua, are capable of producing IL-6 (27). Similarly, kera-
tinocytes, a normal component ofAF due to fetal desquama-
tion, are also capable of producing IL-6 (26). It is likely that
inflamed tissues within the uterus, containing a variety ofdif-
ferent cell types (macrophages, endothelial cells, and fibro-
blasts) produce large amounts of IL-6. The observation that
decidual tissue explants, which contain many of these cell
types, can be induced by LPS to secrete IL-6 is consistent with
this possibility. In situ nucleic acid hybridization analyses for
cells containing IL-6 mRNA and immunohistochemical local-
ization of IL-6 protein (41) in tissue sections are procedures
now available to directly address questions concerning the
sources ofAF IL-6.
The production of IL-6 is a general feature of processes in
which the integrity ofa tissue is challenged. Normal pregnancy
ending in spontaneous labor calls forth major changes in host
biochemistry; intraamniotic infection greatly adds to this
stress. Parturition itselfmay be viewed as part ofa repertoire of
host defense mechanisms elicited by intraamniotic infection.
The production of abundant IL-6 would clearly contribute to
the production of protective acute-phase plasma proteins and
the activation of immune mechanisms that would help limit
tissue damage. IL-6 production at the maternal-fetal interface
may restrict tissue damage to the intrauterine compartment
and protect the mother from the systemic consequences of
disseminated infection. Likewise, the fetus may produce IL-6
in response to localized bacterial invasion (i.e., after aspiration
of infected AF), leading to a protective acute phase plasma
protein response in the fetus. This interpretation is compatible
R. Romero, C. Avila, U. Santhanam, and P. B. Sehgal
with the observation that infected neonates with elevated CRP
have higher survival rates than those with nondetectable serum
CRP (42, 43).
The present study identifies AF IL-6 as a marker ofPTL. A
larger prospective study will be necessary to further substan-
tiate the value of AF IL-6 as a marker cytokine in clinical
practice. From a biological point of view, we propose that
various cytokines may contribute to different aspects of the
host response to intrauterine infection. IL-1 and TNF, which
strongly stimulate prostaglandin biosynthesis by intrauterine
tissues (6, 44), may signal the onset of parturition and also
strongly upregulate IL-6 production. IL-6, in turn, may or-
chestrate biochemical, immunological, and physiological
changes that contribute to maternal and fetal survival.
We thank Mr. Ralph Zinner for excellent technical assistance.
This research was supported by grants from the Walter Scott
Foundation for Medical Research, the National Institutes of Health
(AI-16262), a Physician-Scientist Award from the National Institutes
ofHealth (to Dr. Romero), and a contract from the National Founda-
tion for Cancer Research.
1. van den Berg, B. J., and F. W. Oechsli. 1984. Prematurity. In
Perinatal Epidemiology. M. Bracken, editor. Oxford University Press,
London, UK. 69-85.
2. Romero, R., and M. Mazor. 1988. Infection and preterm labor.
Clin. Obstet. Gynecol. 31:553-584.
3. Romero, R., R. Quintero, E. Oyarzun, Y. K. Wu, M. Mazor, V.
Sabo, and J. C. Hobbins. 1988. Intraamniotic infection and the onset
of labor in preterm rupture of membranes. Am. J. Obstet. Gynecol.
4. Romero, R., M. Sirtori, E. Oyarzun, C. Avila, M. Mazor, R.
Callahan, V. Sabo, A. Athanassiadis, and J. C. Hobbins. 1989. Preva-
lence, microbiology and clinical significance ofintraamniotic infection
in women with preterm labor and intact membranes. Am. J. Obstet.
5. Romero, R., D. T. Brody, Y. K. Wu, M. Mazor, E. Oyarzun,
J. C. Hobbins, and S. Durum. 1989. Infection and labor. III. Interleu-
kin- 1: a signal for the initiation ofparturition. Am. J. Obstet. Gynecol.
6. Romero, R., K. R. Manouge, M. D. Mitchell, Y. K. Wu, E.
Oyarzun, J. C. Hobbins, and A. Cerami. 1989. Infection and labor. IV.
Cachectin-tumor necrosis factor in the amniotic fluid ofwomen with
intraamniotic infection and preterm labor. Am. J. Obstet. Gynecol.
7. Gauldie, J., C. Richards, D. Harnish, P. Lansdorp, and H. Bau-
mann. 1987. Interferon
J2/B-cell stimulatory factor type 2 shares
identity with monocyte-derived hepatocyte stimulating factor and reg-
ulates the major acute phase protein response in liver cells. Proc. Nati.
Acad. Sci. USA. 84:7251-7255.
8. May, L. T., J. Ghrayeb, U. Santhanam, S. B. Tatter, Z. Sthoeger,
D. C. Helfgott, N. Chiorazzi, G. Grieninger, and P. B. Sehgal. 1988.
Synthesis and secretion of multiple forms of (32-interferon/B cell dif-
ferentiation factor-2-hepatocyte stimulating factor by human fibro-
blasts and monocytes. J. Biol. Chem. 263:7760-7766.
9. May, L. T., U. Santhanam, S. B. Tatter, N. Bhardwaj, J.
Ghrayeb, and P. B. Sehgal. 1988. Phosphorylation ofsecreted forms of
human j32-interferon/hepatocyte stimulating factor/interleukin-6.
Biochem. Biophys. Res. Commun. 152:1144-1150.
10. Helfgott, D. C., S. B. Tatter, U. Santhanam, R. H. Clarick, N.
Bhardwaj, L. T. May, and P. B. Sehgal. 1989. Multiple forms of
IFN-l2/IL-6 in serum and body fluids during acute bacterial infection.
J. Immunol. 143:948-953.
11. May, L. T., G. Torcia, F. Cozzolino, A. Ray, S. B. Tatter, U.
Santhanam, P. B. Sehgal, and D. Stern. 1989. Interleukin-6 gene ex-
pression in human endothelial cells: RNA start sites, multiple IL-6
proteins and inhibition ofproliferation. Biochem. Biophys. Res. Com-
12. Jablons, D. M., J. J. Mule, J. K. McIntosh, P. B. Sehgal, L. T.
May, C. M. Huang, S. A. Rosenberg, and M. T. Lotze. 1989. Interleu-
kin-6/interferon-(32 as a circulating hormone: induction by cytokine
administration in man. J. Immunol. 142:1542-1547.
13. Fong, Y., L. L. Moldawer, M. Marano, H. Wei, S. B. Tatter,
R. M. Clarick, U. Santhanam, D. Sherris, L. T. May, P. B. Sehgal, and
S. F. Lowry. 1989. Endotoxemia elicits increased circulating,62-IFN/
IL-6 in man. J. Immunol. 142:2321-2324.
14. Sehgal, P. B., A. Zilberstein, M. R. Ruggieri, L. T. May, A. C.
Fergusson-Smith, D. L. Slate, and F. H. Ruddle. 1986. Human chro-
mosome 7 carries the fl2-interferon gene. Proc. NatL. Acad. Sci. USA.
15. Ferguson-Smith, A. C., Y. F. Chen, M. S. Newman, L. T. May,
P. B. Sehgal, and F. H. Ruddle. 1988. Regional localization of the
interferon-#2/Bcell stimulatory factor 2/hepatocyte stimulating factor
gene to human chromosome 7pl 5-p2 1. Genomics. 2:203-208.
16. Bowcock, A., J. R. Kidd, M. Lathrop, L. Daneshvar, L. T. May,
A. Ray, P. B. Sehgal, K. K. Kidd, and L. L. Cavalli-Sforza. 1988. The
human"interferon-,62/B cell stimulating factor/interleukin-6" gene:
DNA polymorphism studies and localization to chromosome 7p21.
17. Kohase, M., D. Henriksen-DiStefano, L. T. May, J. Vilcek, and
P. B. Sehgal. 1986. Induction of(2-interferon by tumor necrosis factor:
a heomeostatic mechanism in the control of cell proliferation. Cell.
18. Kohase, M., L. T. May, I. Tamm, J. Vilek, and P. B. Sehgal.
1987. A cytokine network in human diploid fibroblasts: interactions of
,3 interferons, tumor necrosis factor, platelet-derived growth factor and
interleukin-l. Mol. Cell. Bio. 7:273-280.
19. Sehgal, P. B., Z. Walther, and L. T. May. 1987. Rapid enhance-
ment of(32-interferon/B cell differentiation factor BSF-2 gene expres-
sion in human fibroblasts by diacylglycerols and the calcium iono-
phore A23187. Proc. Natl. Acad. Sci. USA. 84:3663-3667.
20. Helfgott, D. C., L. T. May, Z. Sthoeger, I. Tamm, and P. B.
Sehgal. 1987. Bacterial lipopolysaccharide (endotoxin) enhances ex-
pression and secretion of (2-interferon by human fibroblasts. J. Exp.
21. Walther, Z., L. T. May, and P. B. Sehgal. 1988. Transcriptional
regulation of the interferon-f2/B cell differentiation factor BSF-2/he-
patocyte stimulating factor HSF gene in human fibroblasts by other
cytokines. J. Immunol. 140:974-977.
22. Sehgal, P. B., D. C. Helfgott, U. Santhanam, S. B. Tatter, R. H.
Clarick, J. Ghrayeb, and L. T. May. 1988. Regulation of the acute
phase and immune responses in viral disease. Enhanced expression of
the "132-interferon/hepatocyte stimulating factor/interleukin-6" gene
in virus-infected human fibroblasts. J. Exp. Med. 167:1951-1956.
23. Zhang, Y., Y. Lin, and J. Vilcek. 1988. Synthesis of interleu-
kin-6(interferon-f2l1B cell stimulatory factor 2) in human fibroblasts is
triggered by an increase in intracellular cyclic AMP. J. Bid. Chem.
24. Ray, A., S. B. Tatter, L. T. May, and P. B. Sehgal. 1988.
Activation of the human "(2-interferon/hepatocyte stimulating fac-
tor/interleukin-6" promoter by cytokines, viruses and second messen-
gers. Proc. Natl. Acad. Sci. USA. 85:6701-6705.
25. Sehgal, P. B., G. Grieninger, and G. Tosato. 1989. Regulation
of the acute phase and immune responses: interleukin-6. Ann. NY
Acad. Sci. 557:1-583.
26. Kupper, T., K. Min, P. B. Sehgal, H. Mizutani, N. Birchall, A.
Ray, and L. May. 1989. Production of IL-6 by keratinocytes: implica-
tions for epidermal inflammation and immunity. Ann. NYAcad. Sci.
Amniotic Fluid Interleukin 6
27. Tabibzadeh, S. S., U. Santhanam, P. B. Sehgal, and L. T. May.
1989. Cytokine-induced production ofinterferon-#2/interleukin-6 by
freshly-explanted human endometrial stromal cells: modulation by
estradiol-17#.J. Immunol. 142:3134-3139.
28. Evans, M. I., S. N. Hajj, L. D. Devoe, N. S. Angerman, and
A. H. Moawad. 1980. C-reactive protein as a predictor of infectious
morbidity with premature rupture of membranes. Am. J. Obstet.
29. Hawrylyshyn, P., P. Bernstein, J. E. Milligan, S. Soldin, A.
Pilard, B. Chir, and F. R. Papsin. 1983. Premature rupture of mem-
branes: the role of C-reactive protein in the prediction of chorioam-
nionitis. Am. J. Obstet. Gynecol. 147:240-246.
30. Potkul, R. K., A. H. Moawad, and K. L. Ponto. 1985. The
association of subclinical infection with preterm labor: the role of
C-reactive protein. Am. J. Obstet. Gynecol. 153:642-645.
31. Dodds, W. G., and J. D. lams. 1987. Maternal C-reactive pro-
tein and preterm labor. J. Reprod. Med. 32:527-530.
32. Guzick, D. S., and K. Winn. 1985. The association ofchorio-
amnionitis with preterm delivery. Obstet. Gynecol. 65:11-16.
33. Moshage, H. J., H. M. J. Roelofs, J. F. van Pelt, B. P. C.
Hazenberg, M. A. van Leeuwen, P. C. Limburg, L. A. Aarden, and
S. H. Yap. 1988. The effect of interleukin-l, interleukin-6 and its
relationship on the synthesis ofserum amyloid A and C-reactive pro-
tein in primary cultures of adult human hepatocytes. Biochem.
Biophys. Res. Commun. 155:112-117.
34. Ganapathi, M. K., L. T. May, D. Schultz, A. Brabenec, J.
Weinstein, P. B. Sehgal, and I. Kushner. 1988. Role ofinterleukin-6 in
regulating synthesis of C-reactive protein and serum amyloid A in
human hepatoma cell lines. Biochem. Biophys. Res. Commun.
35. Caritis, S. N. 1988. A pharmacologic approach to the infusion
of ritodrine. Am. J. Obstet. Gynecol. 158:380-384.
36. Romero, R., K. Scharf, M. Mazor, M. Emamian, J. Ryan, and
J. C. Hobbins. 1988. The clinical value ofgas liquid chromatography
in the detection ofintra-amniotic microbial invasion. Obstet. Gynecol.
37. Tosato, G., K. B. Seamon, N. D. Goldman, P. B. Sehgal, L. T.
May, G. C. Washington, K. D. Jones, and S. E. Pike. 1988. Mono-
cyte-derived human cell growth factor asinterferon-02 (BSF-2, IL-6).
Science (Wash. DC). 239:502-504.
38. Tosato, G., and S. E. Pike. 1988. Interferon-,62/interleukin-6 is
a costimulant for human T lymphocytes. J. Immunol. 141:1556-1562.
39. Kenney, J. S., M. P. Masada, E. M. Eugui, B. M. Delustro,
M. A. Mulkins, and A. C. Allison. 1987. Monoclonal antibodies to
human recombinant interleukin-l (IL-l)#t:quantitation of IL- I3tand
inhibition of biological activity. J. Immunol. 138:4236-4242.
40. Benedetti, T. J. 1983. Maternal complications of parenteral
,B-sympathomimetic therapy for premature labor. Am. J. Obstet. Gy-
41. Grossman, R. M., J. Krueger, D. Yourish, A. Granelli-Piperno,
D. P. Murphy, L. T. May, T. S. Kupper, P. B. Sehgal, and A. B.
Gottlieb. 1989. Interleukin-6 (IL-6) is expressed in high levels in psori-
atic skin and stimulates proliferation ofcultured human keratinocytes.
Proc. Natl. Acad. Sci. USA. 86:6367-6371.
42. Philip, A. G. S. 1985. Response ofC-reactive protein in neona-
tal group B streptococcal infection. Pediatr. Infect. Dis. J. 4:145-148.
43. Philip, A. G. S. 1979. The protective effect of acute phase
response reactants in neonatal sepsis. Acta Paediatr. Scand. 68:481-
44. Romero, R., S. Durum, C. A. Dinarello, E. Oyarzun, J. C.
Hobbins, and M. D. Mitchell. 1989. Interleukin-l stimulates prosta-
glandin biosynthesis by human amnion. Prostaglandins. 37:13-22.
R. Romero, C. Avila, U. Santhanam, and P. B.Sehgal